FULL PAPER
DOI: 10.1002/chem.200((will be filled in by the editorial staff))

“This is the peer reviewed version of the following article: Chem. Eur. J. 2014, 20, 2264-2275, which has been published in final form at
https://doi.org/10.1002/chem.201302536. This article may be used for non-commercial purposes in accordance with Wiley Terms and
Conditions for Self-Archiving published at http://olabout.wiley.com/WileyCDA/Section/id-820227.html."

1

Highly Active Aluminium Catalysts for the Formation of Organic Carbonates from
CO2 and Oxiranes
Christopher J. Whiteoak,[a] Nicola Kielland,[a] Victor Laserna,[a] Fernando Castro-Gómez,[a] Eddy
Martin,[a] Eduardo C. Escudero-Adán,[a] Carles Bo,[a][c] and Arjan W. Kleij*[a,b]
Abstract: AlIII complexes of aminotris-phenolate ligand scaffolds have
been prepared in order to attain highly
Lewis acidic catalysts. Combination of
the aforementioned systems with
ammonium halides provides highly
active catalysts for the synthesis of
organic carbonates through addition of
carbon dioxide to oxiranes with initial
turnover frequencies among the highest

reported to date within the context of
cyclic carbonate formation. Density
functional theory (DFT) studies
combined with kinetic data provides a
rational for the relative high activity
found for these AlIII complexes, and data
are consistent with a monometallic
mechanism. The activity and versatility
of these AlIII complexes has also been
evaluated against some state-of-the-art

catalysts and the combined results
compare favorably in terms of catalyst
construction, stability, activity and
applicability.
Keywords: aluminium • catalysis •
carbon dioxide • organic carbonates
• homogeneous catalysis

Introduction
The use of carbon dioxide (CO2) as a renewable carbon feedstock is
of both academic and industrial relevance.[1] However, there are only
a limited number of chemical processes that have been successfully
developed where CO2 is used as a reagent[2] and even less that classify
as truly sustainable from the point view of energy use and carbon
recycling.[3] One method of increasing the sustainability of a chemical
process is through the use of catalysis and therefore use of such may
enable the realization of truly sustainable CO2 conversions.[4] The
largest challenge in CO2 conversion is its relative high kinetic and
thermodynamic stability; whereas catalysis can be used to effectively
lower the barriers associated with the kinetic stability, the choice for
higher-energy substrates can help to overcome the thermodynamic
hurdle. With this latter aspect in mind, the use of epoxides (oxiranes)
has the advantage that ring-strain is released upon ring-opening and
its conversion in the presence of CO2 providing either organic poly[5] or cyclic carbonates[6] (Scheme 1) is indeed exothermic. This
addition[7] of CO2 to epoxides is a well-documented area of research
where a number of different catalytic solutions have been proposed
using ionic liquids,[8] binary[9] or bifunctional (metallosalen)
complexes,[10] and simple organic motifs such as “onium salts”.[11]
However, only in a limited amount of cases have these catalytic
efforts resulted in systems with (very) high activities expressed in
high turnover frequencies (TOFs) and high turnover numbers
(TONs).

[a]

Dr. C. J. Whiteoak, Dr. N. Kielland, V. Laserna, F. Castro-Gómez, Dr. E.
Martin, E. Escudero-Adán, Prof.C. Bo, Prof. Dr. A. W. Kleij
Institute of Chemical Research of Catalonia (ICIQ)
Av. Països Catalans 16, 43007 Tarragona, Spain
Fax: (+34)-977-920-828
E-mail: akleij@iciq.es

[b]

Prof. Dr. A. W. Kleij
Catalan Institute of Research and Advanced Studies (ICREA)
Pg. Lluís Companys 23, 08010 Barcelona, Spain

[c]

Scheme 1. Top: addition of CO2 to epoxides yielding cyclic- and poly-carbonates and
below some reported successful catalyst systems (A-C) based on aluminium for this
reaction. Note that catalyst systems A and C are binary systems and are combined with
NBu4X (X = Br or I) whereas catalyst B is a one-component, bifunctional system.

Prof. Dr. C. Bo, Departament de Química Física i Inorgánica, Universitat
Rovira i Virgili, Marcel·lí Domingo s/n 43007 Tarragona (Spain)
Supporting information for this article is available on the WWW under
http://www.chemeurj.org/ or from the author.

2

In addition, when taking into account possible application of catalysts
in an industrial setting[12] long-term stability, recycling features,
accessibility, cost-effectiveness, toxicity and catalyst modularity are
features that are of crucial importance.
Recently, Ema, Sakai and co-workers reported on bifunctional
metalloporphyrin catalysts (M = Mg, Zn) that were able to produce
cyclic carbonates based on terminal epoxides with high TOFs of up
to 12,000 h-1 and TONs up to 103,000.[10b] This work inspired us to
design and construct catalyst systems with similar or even improved
potential in this area with a focus on high reactivity and wide
application potential (substrate scope). We became interested in using
amino-tris-phenolate scaffolds[9b,13a] as ligand alternatives for the
widely used salen and/or salans,[14] as these provide different
molecular conformations upon complexation of suitable metal
ions.[13] We envisioned that the presence of fewer donor atoms in the
plane of the metal would be beneficial for those cases where the steric
requirement of the coordinating epoxide substrate would limit
catalytic turnover. As in the majority of cases binary catalyst systems
are employed (i.e., a metal complex combined with a suitable
nucleophile), the nucleophile needed in the ring-opening step of the
coordinated epoxide would have difficulty approaching the substrate
and entering the coordination sphere of the metal. We have recently
reported that FeIII-based amino-tris-phenolate complexes are efficient
catalysts for the formation of a wide series of cyclic carbonates under
virtually ambient conditions (25ºC, pCO2 = 2 bar) and also effectively
mediate the conversion of internal epoxides.[9b] Further to that,
computational analyses[15] on the Zn(salphen) mediated formation of
cyclic carbonates[4c,9e] have indicated that metallosalen catalyst
systems may have limitations concerning the more sterically
demanding substrates. The fact that amino-tris-phenolate complexes
show better catalytic potential is in line with these theoretical studies.
In an attempt to create even more powerful catalyst systems, we
have turned our focus to aluminium (Al) systems, as recent literature
has demonstrated that complexes based on Al can be attractive within
the context of organic carbonate formation (Scheme 1; see systems
AC).[16] Our preliminary results using an Al(amino-tris-phenolate)
complex (Scheme 1, C) demonstrated unprecedented activity (initial
TOFs up to 36,000 h-1 and TONs exceeding 118.000) and substrate
scope;[17] here we will demonstrate that these kind of complexes
based on this earthabundant metal can be easily fine-tuned giving
rise to catalyst systems with activities among the highest reported to
date. Further theoretical, kinetic, stability and structural studies are
also presented that clearly show a combination of attractive properties
for these powerful catalyst systems during CO2 conversion into
valueadded commodities such as cyclic carbonates. Computational
investigations have shed important insight on reasons for the
observation of high activities of these new Al-based catalysts.

4 in good purity although low yield. This ligand is difficult to prepare
selectively as a result of the poor solubility of both the corresponding
amino-bis-phenol and the amino-tris-phenol in the reaction medium.
Upon formation of the amino-tris-phenol ligands, a degradation may
occur towards the amino-bis-phenol as reported by Solomon and coworkers for ligand 1.[21] In the case of 4 this degradation reaction
appears to be rapid as the poor solubility of the amino-bis-phenol
drives the equilibrium towards this undesired product as soon as the
amino-tris-phenol is obtained, thereby limiting the yield of the
desired ligand.
The Al(amino-tris-phenolate) complexes were obtained through
the reaction of the ligand precursors 1-4 with trimethylaluminium
(AlMe3) in THF (Scheme 2). The characterization data for complexes
1, 2 and 4 are consistent with monomeric complexes, whereby a THF
ligand occupies the apical position of a trigonal bipyrimidal
coordination geometry. Complex 7 is presumed to exist as a dimeric
species (as is its Fe(III) analogue)[13] considering its lower solubility
in most solvents. The molecular structure for 8·THF (Figure 1) was
determined by X-ray crystallography from crystals obtained from
THF/hexane. The molecular structures of 5·THF and 6·THF were
reported previously.[22]

Scheme 2. Synthesis of Al(amino-tris-phenolate) complexes 5-8 from precursor ligands
1-4.

Results and Discussion
Synthesis and analysis of complexes: The amino-tris-phenol ligands
1-2 were prepared by a known Mannich condensation reaction,[18]
whereas 3 was prepared in step-wise way.[19] The ligand bearing
chloride substituents (4) has been reported by Kol and co-workers[20]
but the preparation was achieved using a different work-up procedure
to that reported; after refluxing 2,4-dichlorophenol and
hexamethyltetramine (HMTA) together at 110ºC for 2.5 h, an
(unexpected) adduct of HMTA and the desired ligand, 4·HMTA, was
isolated (X-ray structure, see Supporting Information). The HMTA
could be removed by an acid/base work-up to yield the desired ligand

Figure 1. Displacement ellipsoid plot of the molecular structure of 8·THF together with
a partial numbering scheme. Co-crystallized solvent molecules and rotational disorder in
the THF ligand is omitted for clarity. Selected bond distances [Å] and angles [º]: Al(1)O(1) = 1.7604(14), Al(1)-O(2) = 1.7527(15), Al(1)-O(3) = 1.7689(15), Al(1)-O(4) =
1.9216(15), Al(1)-N(1) = 2.0715(17); O(4)-Al(1)-N(1) = 178.11(7), O(1)-Al(1)-O(2) =
120.94(7), O(2)-Al(1)-O(3) = 117.02(7), O(1)-Al(1)-O(4) = 87.31(7), O(2)-Al(1)-O(4) =
89.50(7), N(1)-Al(1)-O(1) = 90.94(7), N(1)-Al(1)-O(3) = 91.43(7).

3

Further fine-tuning of the amino-tris-phenol scaffold can also be
accomplished through the synthesis of nonsymmetrical ligands. In
this respect, amino-bis-phenols (i.e. a secondary amine) may serve as
a suitable precursor, since simple alkylation of the amine with an
appropriate reagent directly affords nonsymmetrical systems. This
was probed for the secondary amine 9[23] which was alkylated with
the commercially available alkyl bromide 10; the desired
nonsymmetrical amino-tris-phenol ligand 11 was isolated in 68%
yield. Metalation, as described for Al-complexes 5-8, was achieved
similarly to furnish 12 in 78% yield. The identity of 12 was first
established by 1H NMR (details in the Supporting Information) and
further affirmed by X-ray crystallography (Figure 2). The structure
reported for 12 represents to our knowledge the first
crystallographically characterized nonsymmetrical amino-trisphenolate complex.

Table 1. Initial catalysis results obtained for the synthesis of the cyclic carbonate derived
from 1,2-epoxyhexane at 30ºC using NBu4I as co-catalyst.[a][b]

Entry

Cat.

Cat.

Co-cat.

Conv.

TOF

(mol%)

(mol%)

(%)[c]

(h-1)[d]

1

0

0

0

5

0.05

0.25

6

60

3

5

0.5

2.5

74

74

4

5

0.7

2.5

94

67

5

6

0.05

0.25

5

50

6

6

0.5

2.5

65

65

7

6

0.7

2.5

83

59

8

7

0.05

0.25

2

20

9

7

0.5

2.5

18

18

10

8

0.5

0

0

0

11

8

0.05

0.25

10

100

12

8

0.5

2.5

91

91

13

12

0.05

0.25

4

40

14

12

0.5

2.5

27

27

15





0.25

0



16





2.5

0



17
Figure 2. Displacement ellipsoid plot of the dimeric structure of 12 together with a partial
numbering scheme. Co-crystallized solvent molecules and rotational disorder are omitted
for clarity. The bond lengths/angles around both Al centres are rather similar. Selected
bond distances [Å] and angles [º]: Al(1)-O(1) = 1.740(3), Al(1)-O(2) = 1.734(3), Al(1)O(3) = 1.846(2), Al(1)-O(8) = 1.948(2), Al(1)-N(1) = 2.003(3); O(1)-Al(1)-O(2) =
124.28(13), O(1)-Al(1)-O(3) = 115.08(12), O(1)-Al(1)-O(8) = 91.15(11), O(1)-Al(1)N(1) = 93.50(12), O(2)-Al(1)-O(8) = 91.14(11), O(3)-Al(1)-O(8) = 76.99(10), Al(1)O(8)-Al(2) = 39.94(8), Al(1)-O(8)-Al(2)-O(3) = 176.71(18).

0.5

2

Scheme 3. Synthesis of nonsymmetrical Al(amino-tris-phenolate) complex 12.

5

D

0.5

2.5

46

46

18

A

0.5

2.5

6

6

19

A

0.5

2.5[e]

50

50

[a] Reaction conditions: 1.0 g 1,2-epoxyhexane (9.98 mmol), 1.0 MPa CO2 initial
pressure, 2 h, 30 ºC. [b] Catalyst concentration calculated on a per Al atom basis and
amount related to the substrate. [c] Determined by 1H NMR spectroscopy of the crude
reaction mixture; selectivity in all cases 99%. Conversions were found to be

Complex 12 crystallized as a dimeric structure comprising two 2phenoxo bridges. Such dimerization potential has been well-

reproducible within ±2% in four independent runs for selected experiments. [d] Average
turn-over frequency (TOF/h) during the first 2 h. [e] Using NBu4Br as co-catalyst.

4

documented in the literature and recently we have shown that dimers
of Fe(amino-tris-phenolate) complexes can be easily disrupted by
addition of a suitable, competing ligand (THF, pyridine) or substrate
(epoxide)[9b,13] providing monomeric species which is a prerequisite
for application in catalysis (vide infra).
Catalysis screening with Al complexes 5-8 and 12: The catalytic
potential of the Al complexes was first evaluated under mild
conditions (T = 30ºC, p(CO2) = 1.0 MPa = 10 bar) using 1,2epoxyhexane (13) as benchmark substrate and NBu4I as co-catalyst
(Table 1, entries 1-14). In general, at catalyst loadings of 0.05 mol%
low conversions were noted, while increasing this by at least ten-fold
(0.50.7 mol%) resulted in high conversion levels of up to 94% (entry
3). The Al-complexes themselves proved to be inactive (entries 1 and
10). The best result, in terms of initial TOF, was obtained for Al
complex 8 (entry 11) which gave an average TOF of 100 h-1. It should
be noted that the co-catalyst alone (NBu4I) under similar conditions
reported for the Al complexes (either at 0.05 or 0.5 mol% catalyst
loading) proved to be ineffective and no conversion into carbonate
was detected (entries 15 and 16). Further to that, upon comparison
with two benchmark catalyst systems (North´s bimetallic Al(salen) A
combined with NBu4I (entry 18) or NBu4Br (entry 19) and our
previously reported Zn(salphen) catalyst D) we found that these latter
systems were less effective and showed about half the activity (TOFs
are 50 h-1 and 46 h-1, respectively; entries 17 and-19) compared with
Al complex 8 at the same metal loading (entry 12; TOF = 91 h-1).[24]
The lowest activities (expressed in initial TOFs) were found for the
non-substituted Al complex 7 and the nonsymmetrical Al complex
12. It is known that these complexes have higher tendency to form
dimeric structures which can compete with the required substrate
coordination. Temperature is an important factor controlling the
monomerdimer equilibrium with higher temperatures favoring
higher concentrations of the more reactive monomeric species.[13]
Therefore, in order to compare and evaluate better the catalytic
potential of Al complexes 5-8 and 12 we then focused on higher
reaction temperatures to further increase the turnover frequencies
(Table 2).
For this second stage of the catalysis screening, temperatures up
to 110ºC were considered and Al complexes 5-8 were utilized as their
substitution pattern allows for a direct comparison of activities within
this series. As expected, at higher temperatures complexes 58
display higher reactivity and an increase in the observed average
TOFs for all complexes, with the complexes in the absence of NBu4I
being inactive (entries 3 and 17; 90ºC). For instance, for Al complex
5 at 50ºC an increase in the TOF is noted to 410 h-1 and reaching 2200
h-1 at 110ºC at a catalyst loading of only 0.005 mol% (Table 2, entries
1-5). For the other catalysts 6-8 similar features were noted, i.e. an
increase in the TOF with the highest reactivity was noted at 110ºC.
The Al complex 7 which is likely to form (at least in part) a dimeric
structure in solution shows, as expected, higher reactivity at higher
temperatures as a result of a more favorable monomer-dimer
equilibrium.[13] Nonetheless, clearly for Al complex 8 at all
temperatures the highest TOFs were observed suggesting that this
system has the highest catalytic potential with a notable TOF of up to
2600 h-1 (entry 20).
At increased reaction temperatures there are two crucial features
to consider: first, at elevated temperatures the efficacy of the cocatalyst alone (i.e., the ammonium salt)[11] is needed for
comparison.[25] Critical assessment of the conversion data at 110ºC
shows that the absence/presence of any Al complex (entries 5, 10,

Table 2. Catalysis results obtained for the synthesis of the cyclic carbonate derived from
1,2-epoxyhexane at various temperatures using NBu4I as co-catalyst.[a][b]

Entry

Cat.

Cat.

Co-cat.

T

Conv.

TOF

(mol%)

(mol%)

(ºC)

(%)[c]

(h-1)[d]

1

5

0.05

0.25

50

41

410

2

5

0.05

0.25

70

60

600

3

5

0.05

0

90

0

0

4

5

0.05

0.25

90

89

890

5

5

0.005

0.025

110

22

2200

6

6

0.05

0.25

50

20

200

7

6

0.05

0.25

70

52

520

8

6

0.05

0.25

90

74

740

9

6

0.01

0.05

90

23

1150

10

6

0.005

0.025

110

17

1700

11

7

0.05

0.25

50

15

150

12

7

0.05

0.25

70

58

580

13

7

0.05

0.25

90

78

780

14

7

0.005

0.025

110

21

2100

15

8

0.05

0.25

50

41

410

16

8

0.05

0.25

70

82

820

17

8

0.05

0

90

0

0

18

8

0.05

0.25

90

96

960

19

8

0.01

0.05

90

42

2100

20

8

0.005

0.025

110

26

2600

21





0.25

50

2



22





0.25

70

15



23





0.25

90

17



24





0.025

110

22



[a] Reaction conditions: 1.0 g 1,2-epoxyhexane (9.98 mmol), 1.0 MPa CO2 initial
pressure, 2 h. [b] Catalyst concentration calculated on a per Al atom basis and amount

5

related to the substrate. [c] Determined by 1H NMR spectroscopy of the crude reaction
mixture; selectivity in all cases 99%. Conversions were found to be reproducible within
±2% in four independent runs for selected experiments. [d] Average turn-over frequency
(TOF/h) during the first 2 h.

14 and 20 versus 24) gives rise to almost the same amount of product
formation and the effect of the presence of the Lewis acidic Al
complex is thus minimal. At 90ºC however, a significant difference
in conversion is still noted (compare for instance entries 19 and 23;
42% versus 17%) in line with the hypothesis that the Al complex
under these conditions plays an imperative role. Second, while up to
a 90ºC reaction temperature the selectivity for the cyclic carbonate is
maintained (99% as determined by 1H NMR), at 110ºC in some
cases we observed trace amounts of sideproducts that probably arise
from thermal decomposition of the carbonate target. Therefore,
because
of
the
above-mentioned
reasons
further
optimization/investigations of the TOF were carried out with Al
complex 8 at a maximum temperature of 90ºC.

Table 3. Optimization studies towards the activity of Al-catalyst 8 in the synthesis of the
cyclic carbonates derived from epoxides 13-16 at 90ºC using NBu4I as co-catalyst and
comparison with reactions using only the co-catalyst; Sub refers to substrate.[a][b]

Entry

Cat. 8

Co-cat.

Sub.

Conv.
[c]

(h-1)[d]

(mol%)

(mol%)

1

Optimization of the performance of Al complex 8 and stability
tests: In order to assess the catalytic potential of Al complex 8 in the
formation of cyclic carbonates, initial TOFs were then determined at
various concentrations of 8 (Table 3). In each case, a comparison was
made between the efficiency of the binary catalyst couple (i.e., Al
complex 8 and NBu4I) versus the co-catalyst alone to ensure that the
Al complex shows a significant contribution in the observed activity.
At a loading of 0.05 mol% (entry 1, 96%) a TOF of 960 h-1 is attained,
whereas the co-catalyst alone (entry 2) shows much lower conversion
(17%). Lowering the loading of Al complex 8 to 0.01 mol% (entry 3)
increases the TOF value to 2100 h-1; it is important to note here that
the actual potential of a catalyst should be measured at the initial stage
of the reaction (i.e., at low conversion levels) when substrate
concentration limitation on the observed activity is minimal. When
decreasing the amount of 8 while maintaining a co-catalyst amount of
0.05 mol% very high initial TOFs of 7600 h -1 (entry 5; 0.0025 mol%
of 8), 16500 h-1 (entry 7; 0.0010 mol% of 8) and 24000 h-1 (entry 8;
0.0005 mol% of 8) were observed whereas the co-catalyst alone
provided a significantly lower conversion (entry 6; 13%). One of the
benchmark systems [Al(salen) dimer A, Scheme 1] from the literature
was again compared with our best performing complex 8. At a total
Al loading of 0.01 mol% (entries 3 and 4 in Table 3), catalyst 8 gave
an initial TOF (2100 h-1) which is four times higher than observed for
dimeric Al(salen) complex A (550 h-1). Other substrates (14-16) were
then tested with Al complex 8 to see whether such high initial TOFs
could also be achieved as observed for 1,2-epoxyhexane 13 (entries
9-11). With these alternative epoxides 14-16 initial TOFs amounted
to 27000 h-1 (14), 24000 h-1 (15) and 17000 h-1 (16), respectively,
demonstrating the wider applicability of catalyst 8 and a very high
activity in general. As a final part of the optimization of the activity
of the binary catalyst couple (Al complex 8 and a suitable co-catalyst)
we then focused on the use of PPN salts [PPN = bis(triphenylphosphoranylidene)ammonium] of bromide and iodide as
these have been shown to give excellent results in the context of
(cyclic/poly) carbonate formation.[26] Whereas the use of Al complex
8/PPN-I (entry 12) showed a modest increase in initial TOF to 29000
h-1 (cf., 24000 h-1 in the case where NBu4I was used; entry 8), the use
of PPN-I alone (entry 13) showed slightly lower conversion of the
substrate. The combination of Al complex 8 and PPN-Br, however,
showed significantly higher activity than the co-catalyst alone
(compare entries 14 and 15) and a remarkably high initial TOF of
36000 h-1 was achieved.

(%)

TOF

0.05

0.25

13

96

960

2



0.25

13

17



3

0.01

0.05

13

42

2100

4

0.01[e]

0.05

13

11

550

5

0.0025

0.05

13

38

7600

6



0.05

13

13



7

0.0010

0.05

13

33

16500

8

0.0005

0.05

13

24

24000

9

0.0005

0.05

14

27

27000

10

0.0005

0.05

15

24

24000

11

0.0005

0.05

16

17

17000

12

0.0005

0.05[f]

13

29

29000

13



0.05[f]

13

21



14

0.0005

0.05[g]

13

36

36000

15



0.05[g]

13

19



[a] Reaction conditions: epoxide substrate (10 mmol), 1.0 MPa CO 2 initial pressure, 2 h,
90ºC. [b] Catalyst concentration calculated on a per Al atom basis and amount related to
the substrate. [c] Determined by 1H NMR spectroscopy of the crude reaction mixture;
selectivity in all cases 99%. Conversions were found to be reproducible within ±2% in
four independent runs for selected experiments. [d] Average turn-over frequency (TOF/h)
during the first 2 h. [e] Using the North catalyst system A from Scheme 1. [f] Using PPNI as co-catalyst. [g] Using PPN-Br as co-catalyst.

In order to establish the robustness of Al complex 8 several
further experiments were conducted. First, at a loading of 0.0005
mol% of 8 using the conditions reported in entry 8 (Table 3) the total
turnover number (TON) was determined after 18 h being a notable

6

112000. This shows that the catalyst is stable enough in the course of
time at 90ºC providing further turnover when compared with the
initial stage of the reaction (entry 8; 2 h).

Table 4. Reactivity in time and stability studies using Al-catalyst 8 in the synthesis of
cyclic carbonates derived from epoxides 13 and 15-17 at 70-90ºC using NBu4I as cocatalyst and comparison with reactions using only the co-catalyst.[a][b]

Entry

Cat. 8

Co-cat.

(mol%)

(mol%)

1

0.0005

0.05

2

0.0005

3

Substr.

Time

Conv.

TOF

(h)

(%)[c]

(h-1)[d]

13

0.5

6

24000

0.05

13

1

13

26000

0.0005

0.05

13

1.5

18

0.0005

0.05

13

2

24

0.0005

0.05

13

3

26

17333

6

0.0005

0.05

13

4

30

15000

7

0.0005

0.05

13

5

34

13600

8

0.0005

0.05

13

6

38

12666

9

0.05

0.25

15

18[e]

99(98)[f]



10

0.05

0.25

15

18[g]

84(79)[f]



11

0.05

0.25

16

18[e]

99(92)[f]



12

0.05

0.25

16

18[g]

99(93)[f]



13

0.05

0.25

17

18[e]

99(93)[f]



14

0.05

0.25

17

18[g]

91(86)[f]



Mechanistic considerations: As the mechanism for dimeric
Al(salen) complexes has been recently investigated by North and coworkers,[9d] we have further focused on some of the features of the
operative mechanism[27] when using Al complexes based on aminotris-phenolate scaffolds. Previously, the first step in the Zn(salphen)
catalyzed formation of cyclic carbonates was proposed to be the
coordination of the epoxide to the metal center.[9e,28] We were able to
grow crystals of a similar structure based on Al complex 8 and
oxetane, and its structure is presented in Figure 3.

24000

5

Interestingly, the results from entries 9-14 in Table 4 show that
neither the presence of air nor water has a dramatic long-term effect
(cf., TON) on the conversion of the epoxide nor the isolated yield of
the carbonate product with typical yields under these conditions being
high (up to 98%). Therefore, by combining the results from Tables 14 it may be concluded that Al complex 8 is both a highly active as
well as robust system that upon combination with a suitable
nucleophile (NBu4X, PPN-X; X = Br, I) gives rise to a binary catalyst
couple with initial activities amongst the highest reported to date.

24000

4

reaction conditions, i.e. in the presence of air or water (Table 4) and
comparing the conversion/yield for three different substrates with
those reported previously.[17]

[a] Reaction conditions: epoxide substrate (10 mmol), 1.0 MPa CO 2 initial pressure, 2 h,
90ºC. [b] Catalyst concentration calculated on a per Al atom basis and amount related to
the substrate. [c] Determined by 1H NMR spectroscopy of the crude reaction mixture;

Figure 3. Displacement ellipsoid plot of the structure of 8·oxetane together with a partial
numbering scheme. Co-crystallized solvent molecules are omitted for clarity, and only
one of the crystallographically independent molecules in the unit cell is shown. Selected
bond distances [Å] and angles [º]: Al(1B)-O(1B) = 1.742(2), Al(1B)-O(2B) = 1.759(2),
Al(1B)-O(3B) = 1.768(2), Al(1B)-O(4B) = 1.907(3), Al(1B)-N(1B) = 2.073(3); O(1B)Al(1B)-O(2B) = 124.84(12), O(1B)-Al(1B)-O(3B) = 115.79(11), O(2B)-Al(1B)-O(4B) =
84.56(10), O(1B)-Al(1B)-O(4B) = 88.67(11), O(1B)-Al(1B)-N(1B) = 92.11(12), O(4B)Al(1B)-N(1B) = 175.76(10).

selectivity in all cases 99%. Conversions were found to be reproducible within ±2% in
four independent runs for selected experiments. [d] Average turn-over frequency (TOF/h)
during the first 2 h. [e] At p(CO2) = 1 MPa together with p(Air) = 0.2 MPa at 70ºC. [f] In
brackets isolated yield. [g] Additionally added 10 L (5 mol%) of water, reaction
performed at 70ºC.

Following the average TOF in time (Table 4, entries 1-8) the
TOF/h stays stable within the first two hours of the reaction and
amounts to 12666 h-1 after 6 h; this lower value may be expected when
the kinetics of a process follows a non-zero order dependency on the
substrate(s) concentration. Then, the stability of the catalyst system
was further probed by looking at the turnover under more competitive

7

Figure 5. Free-energy profiles for the ring-expansion addition of CO2 to propylene oxide catalyzed by Zn(salphen)/NBu4I (in blue) and Al-complex 8/NBu4I (in red) considering the
 pathway (i.e., initial ring-opening occurs at the non-substituted carbon centre). As solvent MEK was considered at 25ºC. IC stands for initial complex formation; FC stands for the
intermediate complex having the carbonate product coordinated.

ln(rate)
-7,5
-3

-2,8

-2,6

-2,4

-2,2

-2

-1,8

-1,6

-1,4

-1,2

-1
-8

y = 2,10x - 4,83
R² = 0,98

-8,5

-9

-9,5

ln[Al]
-10

30ºC was chosen in order to isolate the activity of the binary catalyst
system from that of the background conversion related to the cocatalyst which occurs at higher temperatures; the iodide nucleophile
itself (at 0.5 mol%) does not lead to any observable conversion at this
temperature (separately checked; 0% conversion). Activity is only
observed at 30ºC when both the Al-complex 8 and the iodide
nucleophile are present; the absence of either component results in
non-observable formation of product. The results of the kinetic
studies are highlighted in Figure 4. The conversion of the substrate
was determined by 1H NMR analysis after 2 h in all experiments by
analysing crude reaction mixture aliquots. Interestingly, these kinetic
studies appear to reveal a pseudo second order dependence (2.1;
Figure 4) on the (binary) catalyst system. The rate law for this process
may be written as (equation 1):

-10,5

Rate = k·[CO2]a·[epoxide]b·[Al]c·[I]d

(1)

-11

Figure 4. Double logarithmic plot of the order determination in binary catalyst under
pseudo steady state conditions using 1,2-epoxyhexane/CO2 as substrates and different
concentrations of the Al-complex 8 at 30ºC.

The structure for 8·oxetane simply serves to demonstrate a likely
starting point in the conversion of any substrate converted by the
Al(amino-tris-phenolate) complexes reported in this work. Kinetic
studies (see experimental section for more details) were performed
with the objective to examine the order in the binary catalyst. These
kinetic studies were performed with Al complex 8 in the range
0.0630.25 mol% (the range wherein most of the experiments were
carried out with complexes 5-8 and 12) with respect to the substrate
1,2-epoxyhexane 13 at 30ºC and an initial CO2 pressure of 1 MPa (10
bar) with an iodide loading of 0.5 mol%. A reaction temperature of

Here, k represents the rate constant, and [Al] and [I] the concentration
of complex 8 and the iodide, respectively. Assuming steady state
conditions at the beginning of the reaction (and thus CO2 and epoxide
concentrations being more or less constant) equation 1 can be
rewritten as equation 2:
Rate = kobs·[Al]c·[I]d
Rate = kobs·[Cat]e

with kobs = k·[CO2]a·[epoxide]b (2)

with [Cat] = binary catalyst system (3)

Both the Al-complex 8 as well as the iodide nucleophile are
considered to be equally important for catalytic turnover; they do not
act separately and as a result one can assume that c = d and e = c + d.
Therefore equation 2 can be simplified in the form shown in equation
3. By taking the natural logarithm of equation 3, equation 4 is
obtained:

8

ln(Rate) = ln(kobs) + eln[Cat]
(4)
The kinetic data from the double logarithmic plot revealed that e is
close to 2.0 suggesting a (pseudo) second order dependence on the
binary catalyst system. As the rate-determining step likely involves
the Al-complex 8 as well as the iodide, this suggests that only one
metal center is actively involved in the mechanistic cycle in contrast
with the cycle previously reported for dinuclear Al(salen)
complexes.[9d] Also, the computational investigations appear to
confirm the requirement of a single Al-complex for effective catalytic
turnover as will be discussed below in further detail.
In order to obtain further details about the individual reaction
steps and their energy requirements, computational studies were
performed using DFT based methods and considering a monometallic
mechanism. We recently reported on the computational analysis of
the full reaction mechanism based on the binary catalyst Zn(salphen)
D/NBu4I examining various epoxide substrates.[15] A number of the
previously examined substrates[17] converted by Al-complex 8
required methyl ethyl ketone (MEK) as co-solvent, while the former
DFT studies (BP86) were carried out with the Zn(salphen) based
catalyst using DCM as medium. Therefore, we have now recalculated the free energies of all transition states (TSs) and
intermediates that are involved in the Zn(salphen)/NBu4I catalysed
reaction using MEK as medium, and have compared these
computational results with those obtained for the binary catalyst
system 8/NBu4I using propylene oxide at 25ºC at the B3LYP/6311G**/LANL2DZ level of theory (see Figure 5 including schematic
representations of the involved intermediates and transition states).

coordinating O-atom (O2) of the carbonate fragment (Figure 6)
through TS3 giving the final intermediate FC. Herein the transition
states TS3-a and TS3-b also could afford intermediates FC-a and
FC-b, respectively. Thus path-a still continues being the most likely
pathway because of the higher stability of TS3-a over TS3-b by 14.1
kcal·mol-1, and also because the lower absolute value of most other
barriers. In addition, path-a has an activation energy of 34.2 kcal·mol1, with the CO insertion being the rate-limiting step; whereas for
2
path-b, a higher activation energy was found (39.3 kcal·mol-1), and
the rate-limiting step correspond to the ring-closing. Once the
coordinated cyclic carbonate is formed, it is released from the
complex allowing for further epoxide turnover. The overall reaction
is exergonic with a release of 2.4 kcal·mol-1 in the case of both
catalyst systems.

Considering a similar type of monometallic mechanism for both
types of catalysts, the reaction starts with coordination of the epoxide
(propylene oxide) forming an initial complex IC. Then, the iodide
mediates the ring-opening of the coordinated epoxide at the 
position; this step involves the breaking of the CO bond and the
simultaneous formation of a CI bond through TS1 leading to the
metal-alkoxide intermediate Int-1. It can be observed that this process
is energetically more favourable by 21.6 kcal·mol-1 for Al complex 8
compared to the same step involving the Zn(salphen) complex. The
next step is the insertion of the CO2 molecule into the metal-oxygen
bond of intermediate Int-1, leading to the formation of linear
carbonate Int-2 through transition state TS2.
Unlike for the Zn(salphen) catalyst system, for the Al-catalyzed
reaction this process involves two likely pathways (i.e., path-a and
path-b, see Figure 5). Thus, two TS2 structures could be obtained
depending on the relative orientation (cf., axial or equatorial
coordination position of O1 in Figure 6) of the incoming CO2
molecule with respect the metal centre as shown in Figure 6.
Interestingly, in these transition states a hexa-coordinated Al-centre
is observed which helps to decrease the energy requirement of the
CO2 insertion step. Although TS2-a, and TS2-b are lower in energy
than the corresponding transition state for the Zn(salphen) system by
11.3 and 20.2 kcal·mol-1, respectively, TS2-b would be the most
favourable route due to the lower barrier required to afford Int-2b.
However, since axial-to-equatorial inter-conversion between
intermediates Int2-a and Int2-b should be fairly easy, it is proposed
that these species can undergo isomerization. Intermediate Int2-a was
found more stable than Int2-b by 7 kcal·mol-1, thus suggesting that
the reaction proceeds via path-a. Following the reaction coordinate,
the subsequent step is the ring-closing which occurs via the non-

Figure 6. Optimized structures of the transition states TS1–TS3 for the Al-complex
8/NBu4I catalyzed reaction, together with the most relevant calculated distances (in Å)
and the values of the negative (imaginary) vibrational frequencies.

The competing mechanism was also examined by using the
energetic span (δE) model.[29] Computed free energy spans for path-a
and path-b are fairly high (δE = 34.2 kcal mol-1 and 39.3 kcal mol-1,

9

respectively); these are identical to the activation barrier of each
pathway. IC was found to be the TOF determining intermediate (TDI)
for both catalytic cycles, and TS2-a and TS3-b correspond to the TOF
determining TS (TDTS) of the alternative cycles. On the other hand,
by mixing both pathways (considering the most stable intermediates
and TS), a smaller energetic span of 25.3 kcal mol-1 was obtained,
where IC is the TDI. However, the exact position of the TDTS is
difficult to distinguish because the small difference between TS2-b
(15.4 kcal mol-1) and TS3-a (15.3 kcal mol-1). Consequently, a similar
degree of TOF control of 0.54 for TS2-b and 0.46 for TS3-a is
observed.
The energetic span of the catalytic cycle involving the
Zn(salphen) complex D was found to be higher (δE = 35.6 kcal mol1) compared to the energetic span of the binary Al-containing catalyst
system. The latter, together with the fact that for Al complex 8 all the
involved TSs and intermediates are significantly less energydemanding, is in line with the observation that complex 8 is a much
more efficient catalyst than Zn(salphen) D (see also Table 1 for
experimental evidence). Of particular note is the fact that the initial
coordination of the epoxide in the case of the Zn(salphen) complex is
endothermic (+5.6 kcal·mol-1), while the coordination of propylene
oxide to Al complex 8 is exothermic by 9.9 kcal·mol-1. This result
also demonstrates that 8 is a much more powerful Lewis acid
mediator than Zn(salphen) complex D.

Conclusion
In summary, we have reported on a new family of powerful
Al(amino-tris-phenolate) catalysts for the addition of CO2 to epoxides
with initial activities among the highest reported to date in this area
and turnover numbers exceeding 100000. Furthermore, the most
active catalyst based on Al complex 8 is a highly robust system as
shown by various stability and activity studies. Kinetic data and DFT
calculations appear to support that a monometallic mechanism is
likely operative, with the CO2 insertion step being rate-limiting and
the initial coordination of the epoxide to the Lewis acidic centre more
favourable by around 15 kcal·mol-1 compared to our previous
reported Zn(salphen)/NBu4I binary catalyst system. The potential of
Al complex 8 to form a less energy-demanding hexa-coordinated
transition state is a unique feature (cf., DFT studies) together with its
associated lower energetic span are features that help to explain the
higher catalytic potential of this system versus the catalyst based on
Zn(salphen) D. The simple access to Al complex 8 and its high
potential as a strong Lewis acid mediator make this system a highly
potent system for other types of Lewis acid mediated transformations
such as the formation of polycarbonates under appropriate conditions.
Studies towards this objective are currently ongoing in our laboratory.

Experimental Section
All epoxide substrates and reagents were commercially available
and were used as received. Carbon dioxide was purchased from
PRAXAIR and used without further purification. Solvents used in the
synthesis of the complexes were dried using an Innovative
Technology PURE SOLV solvent purification system. The ligands
1[18], 2[18], 3[19] and 4[20] were prepared as described previously
although 4 was isolated by a modified procedure. The synthesis of
Al(amino-tris-phenolate) complexes was carried out using standard
vacuum line and Schlenk or cannular techniques and once
synthesized the products were stored in a vial kept in air. 1H NMR

spectra were recorded on a Bruker AV-400 or AV-500 spectrometer
and referenced to the residual NMR solvent signals. Elemental
analysis was performed by the Unidád de Análisis Elemental at the
Universidad de Santiago de Compostela. Mass spectrometric analysis
and X-ray diffraction studies were performed by the Research
Support Group at the ICIQ.
Synthesis of ligand precursor (4): 2,4-dichlorophenol (4.0 g,
24.5 mmol) and hexamethylene tetramine (HMTA; 1.2 g, 8.5 mmol)
were heated at 110°C for 2.5 h. Cold MeOH (20 mL) was then added
to the yellow solution before being kept at 0°C for 30 minutes. The
solution was sonicated and the precipitate which subsequently formed
was filtered off and dried under vacuum. The white solid was found
to be a co-crystal with formula 4·HMTA (790 mg, 14%). Crystals
suitable for single crystal Xray diffraction studies of this cocrystallized adduct were obtained by storing the MeOH solution at
30 °C for a further 3 days. 1H NMR (400 MHz, 298 K, CDCl3): δ =
7.23 (d, 3JHH = 2.3 Hz, 3H, ArH), 7.01 (d, 3JHH = 2.3 Hz, 3H, ArH),
6.92 (br s, 3H, ArOH), 4.80 (s, 12H, NCH2N), 3.69 (s, 6H, ArCH2N);
13C{1H} NMR (125 MHz, 298 K, CDCl ): δ = 150.50, 129.03,
3
128.63, 125.22, 124.19, 121.16 (all aromatic C), 74.34 (NCH2N),
55.59 (ArCH2N); elemental analysis calcd (%) for C28H29N5O3Cl6: C
47.53, H 3.99, N 10.27; found: C 47.33, H 3.86, N 10.39. The
resulting 4·HMTA compound was then suspended in water (20 mL)
and acidified to pH = 2 with HCl (37% in water) whilst stirring before
being basified to pH = 11, whereby the compound was fully
dissolved. The solution was neutralized using HCl, at which point the
product precipitated. It was then filtered and dried under vacuum. The
resulting yellowish solid was dissolved in a small amount of CH2Cl2
and hexane added to precipitate the product which was finally filtered
and dried to yield a white powder (680 mg, 37 %). 1H NMR (400
MHz, 298 K, CDCl3): δ = 7.24 (d, 3JHH = 2.4 Hz, 3H, ArH), 7.04 (d,
3J
HH = 2.4 Hz, 3H, ArH), 5.53 (br s, 3H, ArOH), 3.73 (s, 6H,
ArCH2N); 13C{1H} NMR (125 MHz, 298 K, CDCl3): δ = 149.83,
129.10, 128.54, 124.84, 124.72, 121.09 (all aromatic C), 55.35
(ArCH2N); MS (ESI, MeOH): m/z (calcd): 540.9 [M]+; found: 541.
elemental analysis calcd (%) for C21H15NO3Cl6: C 46.53, H 2.79, N
2.58; found: C 46.73, H 2.73, N 2.57.
Synthesis of Al complex (5):[22a] To a solution of ligand
precursor 1 (1.0 g, 2.38 mmol) in THF (20 mL) was slowly added
AlMe3 (2M in heptane, 1.19 mL, 2.38 mmol). The solution was
stirred at r.t. for 2 h and then concentrated. Hexane was added to the
concentrate resulting in precipitation of the complex which was
isolated by filtration and further dried under vacuum to yield a white
powder (963 mg, 78%). 1H NMR (400 MHz, 298 K, CDCl3): δ = 6.90
(d, 3JHH = 1.4 Hz, 3H, ArH), 6.63 (d, 3JHH = 1.4 Hz, 3H, ArH), 4.59
(br s, 4H, THF), 4.28 (br d, 2JHH = 13.5 Hz, 3H, ArCH2N), 2.86 (br d,
2J
HH = 13.5 Hz, 3H, ArCH2N), 2.22 (s, 9H, ArCH3), 2.21 (s, 9H,
ArCH3), 2.18 (br s, 4H, THF); 13C{1H} NMR (125 MHz, 298 K,
CDCl3): δ = 154.29, 131.02, 126.97, 126.89, 125.83, 120.63 (all
aromatic C), 71.31 (THF), 58.67 (ArCH2N), 25.60 (THF), 20.40
(ArCH3), 16.98 (ArCH3); MS (MALDI+, dctb): m/z (calcd): 443.2
[M  THF]+; found: 443.
Synthesis of Al complex (6):[22b] This compound was prepared
in a similar manner to 5 by treatment of ligand precursor 2 (1.0 g, 1.49
mmol) with AlMe3 (2M in heptane, 0.75 mL, 1.49 mmol) to yield a
white powder (869 mg, 76%). 1H NMR (400 MHz, 298 K, CDCl3): δ
= 7.25 (d, 3JHH = 2.5 Hz, 3H, ArH), 6.87 (d, 3JHH = 2.5 Hz, 3H, ArH),
4.36 (br s, 1H, THF), 4.33 (br s, 3H, THF ArCH2N), 4.27 (br s, 3H,
THF), 2.95 (br s, 3H, ArCH2N), 2.06 (m, 4H, THF), 1.43 (s, 27H,

10

ArC(CH3)3), 1.29 (s, 27H, ArC(CH3)3); 13C{1H} NMR (125 MHz,
298 K, CDCl3): δ = 154.82, 139.37, 137.11, 123.78, 123.60, 121.43
(all aromatic C), 70.26 (THF), 58.83 (ArCH2N), 34.88 (ArC(CH3)3),
34.03 (ArC(CH3)3), 31.71 (ArC(CH3)3), 29.70 (ArC(CH3)3), 25.42
(THF); MS (MALDI+, dctb): m/z (calcd): 695.5 [M  THF]+; found:
695.
Synthesis of Al complex (7): This compound was prepared in a
similar manner to 5 by treatment of ligand precursor 3 (350 mg, 1.04
mmol) with AlMe3 (2M in heptane, 0.51 mL, 1.04 mmol) to yield a
white powder (286 mg, 76%). 1H NMR (400 MHz, 298 K,
[D5]pyridine): δ = 7.37 (ddd, 3JHH = 7.6 and 7.7, 4JHH = 1.5 Hz, 3H,
ArH), 7.18 (dd, 3JHH = 7.6, 4JHH = 1.5 Hz, 3H, ArH), 7.06 (dd, 3JHH =
7.7, 4JHH = 0.9 Hz, 3H, ArH), 6.92 (ddd, 3JHH = 7.6 and 7.7, 4JHH =
0.9 Hz, 3H, ArH), 4.49 (br s, 3H, ArCH2N), 3.14 (br s, 3H, ArCH2N);
13C{1H} NMR (125 MHz, 298 K, [D ]pyridine): δ = 160.21, 130.77,
5
130.52, 123.50, 120.69, 119.09 (all aromatic C), 59.46 (ArCH2N).
MS (MALDI+, dctb): m/z (calcd): 359.11 [M]+; found: 359.4;
elemental analysis calcd (%) for C21H18NO3Al·H2O·¼(C6H14): C
67.74, H 5.94, N 3.51; found: C 67.41, H 5.95, N 3.25.
Synthesis of Al complex (8): This compound was prepared in a
similar manner to 5 by treatment of ligand precursor 4 (500 mg, 0.92
mmol) with AlMe3 (2M in heptane, 0.46 mL, 0.92 mmol) to yield a
white powder (350 mg, 60%). Crystals suitable for single crystal Xray diffraction studies were obtained by layering of a concentrated
THF solution of the complex with hexane. 1H NMR (400 MHz, 298
K, CDCl3): δ = 7.33 (d, 3JHH = 2.6 Hz, 3H, ArH), 6.90 (d, 3JHH = 2.6
Hz, 3H, ArH), 4.69 (m, 4H, THF), 4.18 (br d, 2JHH = 13.3 Hz, 3H,
ArCH2N), 2.95 (br d, 2JHH = 13.3 Hz, 3H, ArCH2N), 2.23 (m, 4H,
THF); 13C{1H} NMR (125 MHz, 298 K, CDCl3): δ = 152.46, 129.75,
127.13, 124.68, 123.08, 122.16 (all aromatic C), 72.38 (THF), 58.25
(ArCH2N), 25.48 (THF); MS (MALDI+, dctb): m/z (calcd): 564.9 [M
 THF]+; found: 565; elemental analysis calcd (%) for
C25H20AlCl6NO4: C 47.05, H 3.16, N 2.19; found: C 46.75, H 3.22,
N 2.12.
Synthesis of ligand precursor (11): To a solution of 6,6'(azanediyl-bis(methylene))-bis(2,4-di-tert-butylphenol) (0.80 g, 1.76
mmol) and 2-(bromomethyl)-4-nitrophenol (204 mg, 1.76 mmol) in
THF (20 mL), was added with stirring NEt3 (123 μL, 1.76 mmol).
The reaction mixture was stirred at reflux temperature for 18 h and
filtered after cooling. The volatiles were removed from the filtrate and
the obtained residue was purified by column chromatography
(EtOAc:hexane, 20:80 v/v) to yield a yellow powder (721 mg, 68%).
1H NMR (400 MHz, 298 K, CDCl ): δ = 8.04 (dd, 3J
4
3
HH = 8.9 Hz, JHH
= 2.8 Hz 1H, ArH), 7.96 (d, 4JHH = 2.8 Hz, 1H, ArH), 7.24 (d, 4JHH =
2.4 Hz, 2H, ArH), 6.98 (d, 4JHH = 2.4 Hz, 2H, ArH), 6.86 (d, 3JHH =
8.9 Hz, 1H, ArH), 5.91 (br s, 3H, ArOH), 3.85 (s, 2H, ArCH2N), 3.78
(s, 4H, ArCH2N), 1.41 (s, 18H, ArC(CH3)3), 1.28 (s, 18H,
ArC(CH3)3); 13C{1H} NMR (125 MHz, 298 K, CDCl3): δ = 162.40,
151.31, 142.59, 140.41, 135.92, 126.45, 125.69, 125.62, 124.20,
123.19, 121.68, 116.73 (all aromatic C), 57.68, 56.94 (both ArCH2N),
34.60, 34.19 (both ArC(CH3)3), 31.55, 29.88 (both ArC(CH3)3); MS
(ESI, MeOH): m/z (calcd) 603.4 [M  H]; found: 603; elemental
analysis calcd (%) for C37H52N2O5·½H2O: C 72.40, H 8.70, N 4.56;
found: C 72.57, H 8.63, N 4.47.
Synthesis of Al complex (12): This compound was prepared in a
similar manner to 5 by treatment of 11 (500 mg, 0.82 mmol) with
AlMe3 (2M in heptane, 0.41 mL, 0.82 mmol) to yield a white powder
(422 mg, 78%). Crystals suitable for single crystal X-ray diffraction

studies were obtained by layering of a concentrated THF solution of
the complex with hexane. 1H NMR (400 MHz, 298 K, [D5]pyridine):
δ = 8.27 (dd, 3JHH = 8.9, 4JHH = 2.8 Hz, 1H, ArH), 8.13 (d, 4JHH = 2.8
Hz, 1H, ArH), 7.51 (d, 4JHH = 2.3 Hz, 2H, ArH), 7.12 (br s, 2H, ArH),
7.03 (d, 3JHH = 8.9 Hz, 1H, ArH), 4.15 (br s, 3H, ArCH2N), 3.25 (br
s, 3H, ArCH2N), 1.43 (s, 18H, ArC(CH3)3), 1.34 (s, 18H, ArC(CH3)3);
13C{1H} NMR (125 MHz, 298 K, [D ]pyridine): δ = 165.44, 155.13,
5
139.77, 139.07, 137.48, 135.53, 125.82, 125.75, 124.30, 123.70,
123.53, 119.59 (all aromatic C), 58.83, 56.90 (both ArCH2N), 34.61,
34.11 (both ArC(CH3)3), 31.71, 29.52 (both ArC(CH3)3); m/z (calcd)
628.35 [M]+; found: 628.4; elemental analysis calcd (%) for
C37H49N2O5Al·H2O·½(C6H14): C 69.64, H 8.47, N 3.91; found: C
69.59, H 8.78, N 3.80.
Typical catalysis procedure: 1.0 g of 1,2-epoxyhexane, 0.005
mol% of aluminium complex and 0.025 mol% Bu4NI were charged
into a 30 mL stainless steel autoclave. The autoclave was then
subjected to three cycles of pressurization and depressurization with
CO2 (0.5 MPa, 5 bar), before final stabilization of the pressure to 1.0
MPa (10 bar). The autoclave was sealed and heated to the required
temperature and left stirring for 2 h. An aliquot of the resulting
mixture was taken and conversion analyzed by means of 1H NMR
spectroscopy using CDCl3 as the solvent. The identity of the cyclic
carbonate product was confirmed by comparison to previously
reported literature values. For a photo of the reactor please refer to the
Supporting Information.
Kinetic experiments: The AMTEC vessels were charged with
NBu4I and connected to the hardware. A leak test was first performed
with 1.5 MPa (15 bar) of CO2, with the pressure finally reduced to 0.2
MPa (2 bar). The reactors were filled with 0.4 MPa (4 bar) of CO2,
and degassed to 2 bar, and this cycle was repeated three times. A
solution of the Al catalyst 8 in 1.20 ml of 1,2-epoxyhexane 13 was
injected into the reaction vessel, the pressure was raised to 1.5 MPa
(15 bar), and the temperature was set to 30ºC. The reactions were
stirred for 2 h at 600 rpm. The stirrer was then stopped and degassed.
Hereafter, an aliquot of the resultant reaction mixture was taken and
the conversion was determined by means of 1H NMR spectroscopy
using CDCl3 as the solvent. For a photo of the reactor system please
refer to the Supporting Information.
X-ray diffraction studies: The measured crystals were stable
under atmospheric conditions; nevertheless they were treated under
inert conditions immersed in perfluoropolyether as protecting oil for
manipulation. Data Collection: Measurements were made on a
Bruker-Nonius diffractometer equipped with an APPEX 2 4K CCD
area detector, a FR591 rotating anode with MoK radiation, Montel
mirrors and a Kryoflex low temperature device (T = 173 °C). Fullsphere data collection was used with  and  scans. Programs used:
Data collection Apex2 V2011.3 (Bruker-Nonius 2008), data
reduction Saint + Version 7.60A (Bruker AXS 2008) and absorption
correction SADABS V. 2008-1 (2008). Structure Solution:
SHELXTL Version 6.10 (Sheldrick, 2000) was used. Structure
Refinement: SHELXTL-97-UNIX VERSION.
Crystallographic details for Al-complex 8·THF: Formula
C25H20Cl6NO4Al; Mw = 638.10; crystal size 0.30 × 0.15 × 0.05 mm3;
orthorhombic; space group Pbca; a = 17.2360(19) Å, b = 15.7823(17)
Å, c = 19.047(2) Å;  =  =  = 90.0º; V = 5181.2(10)º; Z = 8; calcd =
1.636 mg/M3; (MoK) = 0.733 mm-1; T = 100(2) K; (min/max) =
2.05/26.25; F(000) = 2592; 93683 reflections collected; 5209 unique
reflections (Rint= 0.0535); absorption correction empirical;

11

refinement method: full-matrix least-squares on F2; data/restraints/
parameters: 5209/86/364; GoF on F2 = 1.074; R1 = 0.0287 and wR2 =
0.0663 [I>2(I) ]; R1 = 0.0404 and wR2 = 0.0728 (all data); largest
diff. peak and hole: 0.333 and -0.305 e3·Å-3. CCDC 902502 contains
the supplementary crystallographic data for this structure. These data
can be obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
Crystallographic details for Al-complex 12: Formula
C95H122N4O10Al2; Mw = 1533.93; crystal size 0.15 × 0.05 × 0.05
mm3; monoclinic; space group P2(1)/c; a = 29.417(3) Å, b =
10.9419(12) Å, c = 30.189(3) Å;  =  = 90º,  = 118.896(4)º; V =
8507.2(17)º; Z = 4; calcd = 1.198 mg/M3; (MoK) = 0.095 mm-1; T
= 100(2) K; (min/max) = 0.79/24.80; F(000) = 3304; 79528
reflections collected; 14140 unique reflections (Rint= 0.1388);
absorption correction empirical; refinement method: full-matrix
least-squares on F2; data/restraints/ parameters: 14140/441/1127;
GoF on F2 = 1.015; R1 = 0.0643 and wR2 = 0.1280 [I>2(I) ]; R1 =
0.1523 and wR2 = 0.1658 (all data); largest diff. peak and hole: 0.546
and -0.554 e3·Å-3. This structure is a toluene solvate and has a tertbutyl group disordered. CCDC 938872 contains the supplementary
crystallographic data for this structure. These data can be obtained
free of charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
Crystallographic details for Al-complex 8·oxetane: Formula
C25H19.75NO4.25Cl6.75Al; Mw = 668.44; crystal size 0.50 × 0.05 × 0.05
mm3; triclinic; space group P-1; a = 13.702(4) Å, b = 14.048(4) Å, c
= 14.784(3) Å;  = 94.689(7),  = 93.252(7),  = 100.374(10)º; V =
2782.5(12)º; Z = 4; calcd = 1.596 mg/M3; (MoK) = 0.757 mm-1; T
= 100(2) K; (min/max) = 1.39/28.23; F(000) = 1354; 21763
reflections collected; 21763 unique reflections (Rint= 0.000);
absorption correction empirical; refinement method: full-matrix
least-squares on F2; data/restraints/ parameters: 21763/72/758; GoF
on F2 = 1.021; R1 = 0.0570 and wR2 = 0.1528 [I>2(I) ]; R1 = 0.0928
and wR2 = 0.1898 (all data); largest diff. peak and hole: 0.881 and
0.940 e3·Å-3. The disorder in co-crystallized solvent molecules and
the coordinated oxetane was done with TWINABS.[30] CCDC 938873
contains the supplementary crystallographic data for this structure.
These data can be obtained free of charge from The Cambridge
Crystallographic
Data
Centre
via
www.ccdc.cam.ac.uk/data_request/cif.

We thank ICIQ, ICREA, the the Spanish Ministry of Science and Innovation (MINECO,
projects CTQ-2011-27385 and CTQ2011-29054-C02-02 and FPI to FCG) and
Generalitat de Catalunya (project 2009-SGR-259) for support. We acknowledge Dr.
Noemí Cabello, Vanessa Martínez, and Sofía Arnal for performing the MS studies.
[1]

a) Carbon Dioxide as Chemical Feedstock (Ed.: M. Aresta), Wiley-VCH,
Weinheim, 2010; b) M. Peters, B. Köhler, W. Kuckshinrichs, W. Leitner, P.
Markewitz, T. E. Müller, ChemSusChem 2011, 4, 1216-1240; c) M. Cokoja, C.
Bruckmeier, B. Rieger, W. A. Herrmann, F. E. Kühn, Angew. Chem. Int. Ed. 2011,
50, 8510-8537; d) T. Sakakura, J. C. Choi, H. Yasuda, Chem. Rev. 2007, 107,
2365-2387; e) R. Martín, A. W. Kleij, ChemSusChem 2011, 4, 1259-1263.

[2]

See for instance: a) G. Pagani, Hydrocarbon Process. 1982, 61, 87-91; b) Refer
also to www.novomer.com for a company focusing on green polymer synthesis
using CO2 as a feedstock.

[3]

a) E. A. Quadrelli, J.-L. Duplan, G. Centi, S. Perathoner, ChemSusChem 2011, 4,
1194-1215; b) G. Centi, G. Iaquaniello, S. Perathoner, ChemSusChem 2011, 4,
1265-1273.

[4]

For some original recent examples: a) C. Das Neves Gomes, O. Jacquet, C.
Villiers, P. Thuéry, M. Ephritikhine, T. Cantat, Angew. Chem. Int. Ed. 2012, 51,
187-190; b) T. Kimura, K. Kamata, N. Mizuno, Angew. Chem. Int. Ed. 2012, 51,
6700-6703; c) A. Decortes, A. W. Kleij, ChemCatChem 2011, 3, 831-834; d) I. I.
F. Boogaerts, S. P. Nolan, J. Am. Chem. Soc. 2010, 132, 8858-8859; e) D. R.
Dalton, T. Rovis, Nat. Chem. 2010, 2, 710-711; f) A. Correa, R. Martín, J. Am.
Chem. Soc. 2009, 131, 15974-15975; g) T. Fujihara, K. Nogi, T. Xu, J. Terao, Y.
Tsuji, J. Am. Chem. Soc. 2012, 134, 9106-9109; h) M. R. Kember, C. K. Williams,
J. Am. Chem. Soc. 2012, 134, 15676-15679; i) S. Wesselbaum U. Hintermair, W.
Leitner, Angew. Chem. Int. Ed. 2012, 51, 8585-8588; j) A. Monassier, V. D'Elia,
M. Cokoja, J. Pelletier, J.-M. Basset, F. E. Kühn, ChemCatChem 2013, 5, 13211324.

[5]

a) D. J. Darensbourg, Inorg. Chem. 2010, 49, 10765-10780; b) D. J. Darensbourg,
Chem. Rev. 2007, 107, 2388-2410; c) M. R. Kember, A. Buchard, C. K. Williams,
Chem. Commun. 2011, 47, 141-163; d) G. W. Coates, D. R. Moore, Angew. Chem.
Int. Ed. 2004, 43, 6618-6639; e) X.-B- Lu, D. J. Darensbourg, Chem. Soc. Rev.
2012, 41, 1462−1484; e) G.-P. Wu, W.-M. Ren, Y. Luo, B. Li, W.-Z. Zhang, X.B. Lu, J. Am. Chem. Soc. 2012, 134, 5682-5688.

[6]

a) M. North, R. Pasquale and C. Young, Green Chem. 2010, 12, 1514-1539; b) A.
Decortes, A. M. Castilla, A. W. Kleij, Angew. Chem. Int. Ed. 2010, 49, 98229837; c) P. P. Pescarmona, M. Taherimehr, Catal. Sci. Technol. 2012, 2, 21692187.

[7]

The often used term “cyclo” addition reaction is in our view incorrect as it
formally refers to a pericyclic chemical conversion between one or more
unsaturated molecules. This combination should then give a cyclic product with
a net reduction of the bond multiplicity.

[8]

a) J. Sun, S.-I. Fujita, M. Arai, J. Organomet. Chem. 2005, 690, 3490-3497; b) Y.
Zhao, C. Yao, G. Chen, Q. Yuan, Green. Chem. 2013, 15, 446-452; c) Z.-Z. Yang,
L.-N. He, C.-X. Miao, S. Chanfreau, Adv. Synt. Catal. 2010, 352, 2233-2240.

[9]

a) R. L. Paddock, S. T. Nguyen, Chem Commun. 2004, 1622; b) C. Whiteoak, E.
Martin, M. Martínez Belmonte, J. Benet-Buchholz, A. W. Kleij, Adv. Synth.
Catal. 2012, 354, 469-476; c) R. M. Haak, A. Decortes, E. C. Escudero-Adán, M.
Martínez Belmonte, E. Martin, J. Benet-Buchholz, A. W. Kleij, Inorg. Chem.
2011, 50, 7934-7936; d) M. North, R. Pasquale, Angew. Chem. Int. Ed. 2009, 48,
2946-2948; e) A. Decortes, M. Martínez Belmonte, J. Benet-Buchholz, A. W.
Kleij, Chem. Commun. 2010, 46, 4580-4582; f) J. Langanke, L. Greiner, W.
Leitner, Green Chem. 2013, 15, 1173-1182.

[10]

a) A. Buchard, M. R. Kember, K. G. Sandeman, C. K. Williams, Chem. Commun.
2011, 47, 212-214; b) T. Ema, Y. Miyazaki, S. Koyama, Y. Yano, T. Sakai, Chem.
Commun. 2012, 48, 4489-4491; c) M. V. Escárcega-Bobadilla, M. Martínez
Belmonte, E. Martin, E. C. Escudero-Adán, A. W. Kleij, Chem. Eur. J. 2013, 19,
2641-2648; d) J. Meléndez, M. North, P. Villuendas, C. Young, Dalton Trans.
2011, 40, 3885-3902; e) M. A. Fuchs, T. A. Zevaco, E. Ember, O. Walter, I. Held,
E. Dinjus, M. Döring, Dalton Trans. 2013, 42, 5322-5329.

[11]

V. Caló, A. Nacci, A. Monopoli, A. Fanizzi, Org. Lett. 2002, 4, 2561-2563.

[12]

For the synthesis of cyclic carbonates using a flue gas based feedstock: M. North,
B. Wang, and C. Young, Energy Environ. Sci. 2011, 4, 4163-4170.

[13]

a) C. J. Whiteoak, B. Gjoka, E. Martin, M. Martínez Belmonte, E. C. EscuderoAdán, C. Zonta, G. Licini, A. W. Kleij, Inorg. Chem. 2012, 51, 10639-10649; b)
G. Licini, M. Mba, C. Zonta, Dalton Trans. 2009, 5265-5277.

[14]

For some recent reviews on salen chemistry and systems see for example: a) C. J.
Whiteoak, G. Salassa, A. W. Kleij, Chem. Soc. Rev. 2012, 41, 622-631; b) A. W.
Kleij, Chem. Eur. J. 2008, 14, 10520-10529.

The structure for 4·HMTA can be found in the Supporting
Information, CCDC 938871 contains the supplementary
crystallographic data for this structure.
Computational details: All calculations were performed by
using the Gaussian 09 package[31] with the hybrid B3LYP
functional.[32] The standard 6-311G(d,p) basis set was used to
describe the H, C, N and O atoms. The relativistic effective core
pseudo potential LANL2DZ was used, together with its associated
basis set, for Al, Zn, Cl and I atoms. Full geometry optimizations were
performed without constrains. The nature of the stationary points
encountered was characterized either as minima or transition states
by means of harmonic vibrational frequencies analysis. The zeropoint, thermal, and entropy corrections were evaluated to compute
enthalpies and Gibbs free energies (T = 298 K, p = 1 bar). Solvent
effects for MEK have been introduced by using the Polarizable
Continuum Model (PCM).

Acknowledgements

12

[15]

F. Castro-Gómez, G. Salassa, A. W. Kleij, C. Bo, Chem. Eur. J. 2013, 19, 62896298.

[27]

For an early discussion around fundamental mechanistic features in CO 2
chemistry refer to: W. Leitner, Coord. Chem. Rev. 1996, 153, 257-284.

[16]

a) W. Clegg, R. W. Harrington, M. North, R. Pasquale, Chem. Eur. J. 2010, 16,
6828-6843; b) D. Tian, B. Liu, Q. Gan, H. Li, D. J. Darensbourg, ACS Catal.
2012, 2, 2029-2035.

[28]

M. Taherimehr, A. Decortes, S. M. Al-Amsyar, W. Lueangchaichaweng, C. J.
Whiteoak, A. W. Kleij, P. P. Pescarmona, Catal. Sci. Technol. 2012, 2, 22312237.

[17]

C. J. Whiteoak, N. Kielland, V. Laserna, E. C. Escudero-Adán, E. Martin, A. W.
Kleij, J. Am. Chem. Soc. 2013, 155, 1228-1231.

[29]

[18]

a) A. Chandrasekaran, R. O. Day, R. R. Holmes, J. Am. Chem. Soc. 2000, 122,
1066-1072; b) M. Kol, M. Shamis, I. Goldberg, Z. Goldschmidt, S. Alfi, E. HayutSalant, Inorg. Chem. Commun. 2001, 4, 177-179.

a) S. Kozuch, S. Shaik, Acc. Chem. Res. 2011, 44, 101-110; b) A. Uhe, S. Kozuch,
S. Shaik, Comput. Chem. 2011, 32, 978-985; c) S. Kozuch, WIREs Comput Mol
Sci 2012, 2, 795-815; See also: d) Uhe, M. Hölscher, W. Leitner, Chem. Eur. J.
2012, 18, 170-177 for the use of the energetic span model in CO2 chemistry.

[30]

TWINABS version 2008/4, Bruker AXS; R. H. Blessing, Acta Cryst. 1995, A51,
33-38.

L. J. Prins, M. Mba, A. Kolarović, G. Licini, Tetrahedron Lett. 2006, 47, 27352738.

[31]

Gaussian 09, Revision A.02, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci,
G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov,
J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E.
Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K.
Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N.
Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J.
Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C.
Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A.
Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B.
Foresman, J. V. Ortiz, J. Cioslowski, and D. J. Fox, Gaussian, Inc., Wallingford
CT, 2009.

[32]

a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648-5652; b) C. Lee, R. G. Parr, W.
Yang, Phys. Rev. B 1988, 37, 785-789.

[19]

[20]

S. Groysman, I. Goldberg, M. Kol, E. Genizi, Z. Goldschmidt, Adv. Synth. Catal.
2005, 347, 409.

[21]

X. Zhang, D. H. Solomon, Polymer 1998, 39, 405-412.

[22]

a) A. L. Johnson, M. G. Davidson, Y. Pérez, M. D. Jones, N. Merle, P. R. Raithby,
S. P. Richards, Dalton Trans. 2009, 5551-5558; b) J. Zhang, A. Liu, X. Pan, L.
Yao, L. Wang, J. Fang, J. Wu, Inorg. Chem. 2011, 50, 9564-9570. For some
catalytic applications of Lewis acidic Al(amino-tris-phenolates) see: c) S. M.
Raders, J. G. Verkade, J. Org. Chem. 2009, 74, 5417-5428; d) S. M. Raders, J. G.
Verkade, Tetrahedron Lett. 2009, 50, 5317-5321.

[23]

I. S. Belostotskaya, N. L. Komissarova, T. I. Prokof’eva, L. N. Kurkovskaya, V.
B. Vol’eva, Russian J. Org. Chem. 2005, 41, 703-706.

[24]

Note that a comparison between catalysts at higher conversion levels does not
necessarily imply the best evaluation of the relative catalyst potential. This should
be done at the early stage of the reactions (i.e., at low conversions) where initial
TOFs give a better description of the potential of a catalyst.

[25]

C. Beattie, M. North, P. Villuendas, C. Young, J. Org. Chem. 2013, 78, 419-426.

[26]

Received: ((will be filled in by the editorial staff))
Revised: ((will be filled in by the editorial staff))
Published online: ((will be filled in by the editorial staff))

For illustrative examples: a) W.-M. Ren, G.-P. Wu, F. Lin, J.-Y. Jiang, C. Liu, Y.
Luo, X.-B. Lu, Chem. Sci. 2012, 3, 2094-2102; b) A. Berkessel, M. Brandenburg,
Org. Lett. 2006, 8, 4401-4404; c) D. J. Darensbourg, S. J. Wilson, J. Am. Chem.
Soc. 2011, 133, 18610-18613.

13

Entry for the Table of Contents

Strong Lewis Acids 
Christopher J. Whiteoak, Nicola
Kielland, Victor Laserna, Fernando
Castro-Gómez, Eddy Martin,
Eduardo C. Escudero-Adán, Carles
Bo, and Arjan W. Kleij*……...……
Page – Page
Highly Active Aluminium
Catalysts for the Ring-Expansion
Addition of CO2 to Oxiranes

AlIII derived complexes from readily
accessible amino tris-phenolate
ligands show exceptional catalytic
reactivity in the addition of CO2 to
epoxides
affording
cyclic
carbonates with high initial turnover
frequencies of up to 36,000 h-1 and
TONs exceeding 100,000. A
combined
experimental
and
computational approach has been
used to unveil the key features of
these highly robust and attractive
catalysts in the context of organic
carbonate formation.

14